Assembly System DNA SEQUENCE AND ORGANIZATION OF THE CYTOCHROME b GENE IN SACCHAROMYCES CEREVISIAE D273-10B*

The mitochondrial genomes of cytoplasmic “petite” (p-) mutants of Saccharomyces cerevisiae have been used to sequence the cytochrome b gene. A continuous sequence of 6.2 kilobase pairs has been obtained from 71.4 to 80.2 units of the wild type map. This region contains all the cytochrome b mutations previously assigned to the cobl and cob2 genetic loci. Analysis of the DNA sequence has revealed that in the strain D273- 10B, the cytochrome b gene is composed of three exons. The longest exon (bl) codes for the first 252 to 253 amino acids from the NHz-terminal end of the protein. The next two exons (b2 and b3) code for 16 to 18 and 115 to 116 amino acids, respectively. The complete cy- tochrome b polypeptide chain consists of 385 amino acids. Based on the amino acid composition, the yeast protein has a molecular weight of 44,000. The three exon regions of the cytochrome b gene are separated by two introns. The intron between bl and b2 is 1414 nucleotides long and contains a reading frame that is continuous with the reading frame of exon bl. This intron sequence is potentially capable of coding for another protein of 384 amino acid residues. The second intron is 733 nucleotides long. This sequence is rich in A+T and includes a G+C cluster that may be involved in processing of the cytochrome b messenger. The organization of the cytochrome b region in S. cerevisiae D273-10B is somewhat less complex than has been reported for other yeast strains in which exon b l appears to be further fragmented into

The longest exon (bl) codes for the first 252 to 253 amino acids from the NHz-terminal end of the protein. The next two exons (b2 and b3) code for 16 to 18 and 115 to 116 amino acids, respectively. The complete cytochrome b polypeptide chain consists of 385 amino acids. Based on the amino acid composition, the yeast protein has a molecular weight of 44,000. The three exon regions of the cytochrome b gene are separated by two introns. The intron between b l and b2 is 1414 nucleotides long and contains a reading frame that is continuous with the reading frame of exon bl. This intron sequence is potentially capable of coding for another protein of 384 amino acid residues. The second intron is 733 nucleotides long. This sequence is rich in A+T and includes a G+C cluster that may be involved in processing of the cytochrome b messenger. The organization of the cytochrome b region in S. cerevisiae D273-10B is somewhat less complex than has been reported for other yeast strains in which exon b l appears to be further fragmented into three smaller exons.
In the preceding paper, we reported the restriction map of the region of yeast mitochondrial DNA (mtDNA) that codes for apocytochrome b. The restriction map was constructed from an analysis of the mtDNA segments retained in p -clones with genetic markers in the cob1 and cob2 loci.' The p -clones spanned the wild type genome from 69.3 to 80.2 units and encompassed all the cytochrome b markers tested.
* This research was supported by Grant HL22174 from the National Institutes of Health, United States Public Health Service. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' The abbreviations used are: p -refers to cytoplasmic petite mutants resulting from long deletions in mitochondrial DNA. Mit loci are markers in mitochondrial genes that code for subunit polypeptides of cytochrome oxidase, cytochrome b (cobl and cob2), and the oligomycin-sensitive ATPase. The antibiotic resistance loci referred to in this paper are par (paromomycin), cap (chloramphenicol), and oli (oligomycin). bp, base pair.
The availability of p -mutants with well characterized genomes has made it possible to sequence the cytochrome b region and deduce the primary structure of the protein. Our studies show that in the D273-10B strain of Saccharomyces cereuiszae, the cytochrome b gene is constituted of three exons. The coding and intervening sequences have been identified from the homology of the amino acid sequences of the yeast and mammalian cytochrome b.

MATERIALS AND METHODS
Strains-The isolation and physical characterization of the pstrains have been reported (1).
Purification of mtDNA-The p -mutants were grown at 30°C in 10 liters of liquid YPD medium (2% glucose, 1% yeast extract, and 2% peptone). The cells were harvested in early stationary phase of growth. Mitochondria were isolated and mtDNA was purified as described previously (1).
DNA Sequencing-Restriction fragments were labeled at the 5' ends with [y-32P]ATP (2,000 to 3,000 Ci/mmol, New England Nuclear) in the presence of T4 polynucleotide kinase (2). In most instances, the labeled fragments were denatured either in 0.1 N NaOH or 80% formamide and separated into single strands on polyacrylamide gels (2). In some cases, the labeled fragments were subjected to a second endonucleolytic cleavage before separation of the double-stranded DNA fragments on polyacrylamide gels. The 5' end-labeled fragments were sequenced by the chemical derivatization method of Maxam and Gilbert (2). In most experiments, the single strands were clearly separated without evidence of significant cross-contamination. The absence of any appreciable background radioactivity on the strand separation gels indicated that there were few if any single-stranded nicks in the restriction fragments.

RESULTS
Sequencing Strategy-The p-clones used for the sequencing of mtDNA originated from the respiratory-competent haploid S. cereuisiae D273-10B/A21 (3). The most complex clone DS400/A12 has a mitochondrial genome with a unit length of 7.6 kbp spanning the region of wild type mtDNA from 69.3 to 80.2 units (1). All the other p -clones were derived from DS400/A12. The unit lengths of their mtDNAs ranged from 0.16 to 3.5 kbp (1).
Although partial sequence data was obtained from preparative restriction fragments of the DS400/A12 mtDNA, the genome size of this clone was much too large to be sequenced directly. Our strategy, therefore, was to sequence the mtDNA segments of the simpler p-mutants. Due to the smaller genome sizes, there were fewer restriction sites and it was possible to find combinations of enzymes that generated a suffkiently discrete number of fragments for adequate resolution of 5' end-labeled mixtures by electrophoresis on polyacrylamide gels. In addition, the smaller genome sizes allowed for a greater extent of amplification of the retained segments of mtDNA.
Most of the sequence was determined from the mtDNAs of DS400/M8, DS400/M4, and DS400/Mll. Some remaining This information was important in identifying the coding elements in the cytochrome b region.
Nucleotide Sequence of DS400/A12-The spans of mtDNA represented by the p-clones used for the sequence analysis are shown in Fig. 1. AU the mtDNA segments have been aligned with the restriction map of DS400/A12. The fragments sequenced from each genome are indicated by the arrows. The sequence between 71.4 and 74.3 units was determined primarily from the genomes of DS400/M8 and M4 and the smaller mtDNA segments of DS400/N1 and N24. This region has numerous restriction sites and most of the sequences could be confiied from overlapping fragments and complementary strands.
The middle region of DS400/A12 (74.3 to 75.5 units) was sequenced in the clones DS400/N23, N28, N31, and preparative HincII-Hind111 and Hinff fragments from DS400/A12 (Fig. 1). The sequence of the other half of the mtDNA (75.4 to 80.2 units) was obtained from DS400/Mll and in part from DS400/N2. The sequence of this region has a few short gaps around some of the Hpa I1 sites that could not be crossed from neighboring restriction sites. In addition to the small gaps, only partial sequences were obtained in the region from 69.3 to 71.4 units. Even though we do not have the complete sequence of this region, examination of the sequencing gels The restriction fragments and the approximate distances sequenced are indicated by the arrows. The broken arrows refer to the sequences that crossed the juncture of the mtDNA segments. The HinfI fragments 1, 4, and 5 and the HincII-Hind111 fragment were prepared from the DS400/A12 mtDNA. These preparative pieces were subjected to secondary cleavages with restriction enzymes prior to 5' end labeling. The DS400/N7, N9, and N31 clones whose genomes were also sequenced are not shown on the map.
indicates that it is highly rich in A+T. The only potential coding sequence that could be detected straddles the Tag I/ HinfI/Mob II/Hph I site cluster at 69.9 units. These sites occur in a relatively G+C-rich sequence of 70 to 80 nucleotides that has been identified as the glutamic tRNA gene previously mapped in the general vicinity of the cytochrome b region (4).
The nucleotide sequence from 71.4 to 80.2 units is presented in Fig. 2. A survey of the sequence reveals only three regions with a sufficiently high G+C content and uninterrupted reading frames to have possible coding functions. The fnst region starts with an AUG initiator at nucleotide +1 and continues in frame for 1917 nucleotides before ending with an ochre terminator. The average G+C content of this sequence is 25%. The other two registers of the same DNA strand or the three registers of the complementary strand contain numerous ochre and amber termination codons. The reading frame is preceded by 139 nucleotides consisting almost entirely of A+T The second coding region occurs between nucleotides +2174 and +2227, being only some 50 nucleotides long. The G+C content of this short sequence is 35%. The third region of interest (+2883 to +3308) also has a unique reading frame capable of generating a continuous amino acid sequence of 141 residues. This reading frame terminates with an ochre codon and is followed by an A+T-rich sequence. The G+C content of the coding sequence is 26%.
The amino acid sequences encoded by the three reading frames are indicated above the DNA sequences in Fig. 2 and are based on the usual codon assignments of the universal code with two exceptions. Yeast and human mitochondria have recently been shown to use UGA as a tryptophan codon (5, 6). Yeast mitochondria have also been found to translate CUA as threonine instead of leucine (7). Both deviations from the standard code have been taken into account in deriving the amino acid sequences.
The 2800-nucleotide sequence downstream from the ochre terminator in the third coding region consists of semirepetitive A+T-rich sequences interspersed with numerous G+C clusters. This is particularly evident in the span from 76.2 to 80.2 units where there are nine G+C clusters, each having one or more Hpa I1 and Hue I11 sites.
Structural Gene Coding Elements-The apocytochrome b gene of yeast mitochondria has been shown to have a mosaic structure (8-10). Although cytochrome b has been purified from S. cereuisiae (11,12), Neurospora crassa (13), and bovine (14) mitochondria, none of the proteins have been sequenced. These two circumstances made it impossible to distinguish the exon and intron components of the gene from the DNA sequence alone.
A fortunate solution to this dilemma came from a comparison of the DNA sequences of the yeast and bovine mitochondrial DNAs. Drs. F. Sanger and B. Barrell at the MRC Laboratory in Cambridge, England, have obtained extensive sequence data on bovine mtDNA. Their analysis of the bovine DNA revealed a gene that codes for a protein with amino acid sequence homology to three separate amino acid sequences encoded in the cytochrome b region of the yeast genome. Since the homology of the two proteins is high (approximately 50%), it has permitted a fairly precise definition of the exonintron interfaces in the yeast sequence. It should be stressed, however, that the identification of the coding elements in the yeast sequence is contingent on two assumptions. The first is that there are no common intervening sequences in the yeast and bovine genes coding for homologous amino acid sequences, and the second is that no major deletions or insertions of amino acids exist in the two proteins.
The first homologous amino acid sequence is encoded by (-139 to -1).    +I576 TAT TAT TAT  TCC GGT AAA  TTA AAA ATT AGA CCT CAA TTA ACT ATT AGC GTT ACA  AAT AAA TAT TTA CAT GAT GTT   -H p a I I Ser I l e   +I651 GAA TAC  TAT AGA GAA GTA  TTT GGT GGT  AAT ATT TAT  TTT GAT AAA GCT AAA AAT.GGT  TAT TTT AAA TGA TCT ATT   RK5.r   +I726 AAT AAT AAA GAA TTA CAT AAT  ATT TTT TAT CTT TAT AAT AAA AGT TGT CCT TCT AAA TCT

FIG. 2-Continued
the three exons is shown in Fig. 3. The protein consists of 385 amino acid residues and has a molecular weight of 44,000. There are 2 amino acids at the exon-intron junctures that are not homologous to the bovine cytochrome b. One occurs between bl and b2 and the other occurs between b2 and b3.
If the splicing mechanism does not involve the generation of new codons at the junctures, the alternative amino acids at residue 253 are glutamine or histidine, and at residue 270 are aspartic acid or valine. Localization of Cytochrome b Mutations-The DNA sequence analysis of the low complexity p-clones has made it possible to ascribe precise physical limits to their mtDNA segments and to map certain cytochrome b mutations in the exon and intron regions. The physical limits of the mtDNA segments indicated by the DNA sequence is in excellent agreement with their previous localization in the DS400/A12 mtDNA from restriction analysis (1).
The genotypes of the clones and the nucleotide sequences retained in their genomes are summarized in Table I. Based on the results, the mutations have been assigned to the spans of the DS400/A12 mtDNA shown in Fig. 4. The four cob2 mutations M10-152, M17-71, and M33-119 fall entirely within exon bl. The f i h cob2 marker studied, M9-228, is present in DS400/N7 and N9. The N7 mtDNA contains part of exon bl and in addition extends for 256 nucleotides into the fist intron. Since the mi-mutant M9-228 synthesizes a truncated form of apocytochrome b (15), the mutation is likely to be in exon bl.
The other five markers have previously been mapped in the cob1 locus (15). M8-181 is the only marker retained in the clones DS400/N23 and N28. The genomes of these clones contain the entire exon b2 plus some flanking sequences from the neighboring introns. At present, it is not certain whether M8-181 is an exon or intron mutation. The three mutations

TABLE I Sequences retained in low complexity p -clones
The numbering of the nucleotides is the same as in Fig. 2. Three of the low complexity clones have small internal deletions. Deletions: in DS400/N23, from +2125 to +2174; in DS400/N2, from +3376 to +3397; and in DS400/N7, from -20 to +503. M7-40, M8-53, and M13-101 map in the region of exon b3. The M7-40 marker is retained in DS400/N2 whose genome contains most of exon b3 and a 410-nucleotide sequence downstream from the gene. M7-40 has also been shown to synthesize a prematurely terminated polypeptide (10) and is probably a mutation in exon b3. The M8-53 and M13-101 markers map in a region that includes part of exon b3 and part of the intron separating b2 from b3. These two mutations could be either in exon b3 or in the adjoining intron. The only mutation that clearly lies outside of the exon regions if M6-200. This marker is present in the low complexity clone DS400/N31. The restriction map (1) and DNA sequence of the mtDNA segment in DS400/N31 suggest that it arose from a deletion in the internal region of the parental DS400/M4 mtDNA from 71.3 to 75.2 units. On the linear map, the genome of this clone, therefore, represents two separate spans of the DS400/A12 mtDNA. One lies between nucleo-

DNA Sequence and
Organization of the Cyt b Gene tides +2539 and +2800. The other sequence of DS400/N31 originates from a region upstream from the gene (70.8 to 71.3 units). Since the latter sequence is outside of the gene, the M6-200 mutation must be in the second intron between nucleotides +2539 and +2800. Nucleotide Sequences of the Cytochome b Introns-The three cytochrome b exons define two intron regions. The first intron (bl/b2) is 1414 nucleotides long. The beginning 1158 nucleotides of this intron contain a reading frame that is continuous with the reading frame of exon b l (Fig. 2). There are several interesting features about this sequence. The intron reading frame has an AUG initiation codon 1 or 2 amino acids downstream from the end of exon b2. The reading frame starting with this initiator is 1152 nucleotides long and is capable of generating a protein with 384 amino acids (Fig. 2). The hypothetical protein has an unusual composition with a very high content of lysine, tyrosine, and asparagine. It is also significant that the intron sequence has a G+C content of only 23%. This is somewhat lower than the usual range of 27 to 31% G+C found for yeast mitochondrial genes (16-20).
The second intron (b2/b3) is 733 nucleotides long. Intron b2/b3 contains A+T-rich sequences as well as short stretches of nucleotides with a moderate G+C content. There are no reading frames of sufficient length in this intron to code for a protein. In addition to the semirepetitive A+T sequences, there is a G+C cluster with three closely spaced Hpa I1 sites and a single Hue I11 site (+2769 to +2800). It is of interest that the terminal 14 nucleotides of the Hpa II/Hae I11 site cluster (+2787 to +2800) has the sequence 3"GAGGAA-AGCCCCAA-5'. An inversely homologous sequence, 5" CTCCTTTCGGGGTT-3', occurs some 500 nucleotides upstream in the intron (+2323 to +2335). These sequences have previously been found to be common to many G+C clusters in the yeast genome and have been proposed to serve as recognition sites for processing of mitochondrial transcripts (21). A possible secondary structure of the intron region assuming base pairing between the two inverted repeats is shown in Fig. 5.
Although we have not completed our analysis of the sequences adjacent to the intron-exon interfaces, it should be noted that exons b2 and b3 have a common pentanucleotide sequence, CCTGA, close to their 5' termini. These sequences encode Pro-Asp and Pro-Glu in exons b2 and b3, respectively. To determine whether this short sequence is important in splicing will require additional information about the exact intron-exon junctures and a more detailed analysis of the exon and intron sequences.
Codon Utilization-The cytochrome b gene reflects the same bias in codon utilization as do other yeast mitochondrial genes (16-20). Most of the codons terminate in an A or U and certain codons are not used (Table 11). For example, the CGN family for arginine has not been detected in any of the genes despite the fact that the cognate tRNA gene is present in the yeast genome.2 Certain codons that were not found in the genes sequenced earlier appear in the cytochrome b gene. These are UGC for cysteine and GUG for valine. Table I1 also lists the codons found in the intron coding frame. This frame contains six codons that are absent in the bona fide genes.
The amino acid composition of yeast cytochrome b derived from the DNA sequence is shown in Table 111. The experimentally determined compositions of N. crassa and bovine cytochromes b are also listed for comparison. With the exception of a few amino acids, the compositions of the three different proteins are quite similar.
Juncture of DS400/AlZ"The DS400/A12 genome is a cir-S . Bonitz, Columbia University, unpublished studies.  (1). The restriction map of DS400/A12 indicated that it partially overlaps with the mtDNA segment of another pmutant DS401 (1). Both clones have a common 920-bp Hpa I1 fragment with an internal HinfI site. Since DS401 was selected for the retention of the olil and loss of the cytochrome b markers, its DNA segment lies on the cap side of the cytochrome b gene (21).
In order to determine the juncture nucleotides of the DS400/A12 genome, the sequence of the Hpa 11-HinfI fragments from each of the strains were compared (see Fig. 6 for fragments sequenced). The sequences of the fragments were found to be identical up to 11 nucleotides away from the Hpa I1 site (Fig. 7). The sequence divergence in DS400/A12 starts from nucleotide +6125, indicating that the 11 nucleotides preceding the Hap I1 site are part of a sequence on the other side of the juncture of the DS400/A12 genome. The Hpa I1 site in DS400/A12, therefore, is located at 69.3 map units and is distinct from the Hpa I1 site of the DS401 mtDNA at 80.2 units.
The orientation of the cytochrome b exons relative to the wild type physical map confirms previous conclusions about the direction of transcription of the cytochrome b gene (10, 22). The gene is transcribed from the Same DNA strand as the genes for subunits 6 and 9 of the ATPase (16,17,19) and subunits 2 and 3 of cytochrome oxidase (18,20). The positions of these genes on the physical map of D273-10B and the direction in which they are transcribed are illustrated in Fig.  8.
Organization of the Cytochrome b Gene in 0273-10B and Other Strains of Yeast-Several lines of evidence point to substantial differences in the organization of the cytochrome b gene among laboratory strains of S. cereuisiae. The D273-10B strain used in the present study has a somewhat shorter mitochondrial DNA (70 kbp) than do most other strains examined (75 kbp) (23). The smaller genome size in D273-10B is due in part to a 3.4-kbp deletion in the region of cytochrome b (23). The restriction maps shown in Fig. 9 compare the maps of D273-10B and KL14-4A, a strain whose mitochondrial    (box 5/4, 8, 1/9, 2, and 6)    (box 3, 10, and 7) (8, 10) agree with measurements of the duplex and loop regions observed in electron micrographs of RNA-DNA hybrids (25). The most recent estimates of the sizes of the exons and introns are presented in Table  IV. These data suggest that the bl exon of D273-10B is split into three smaller exons in KL14-4A and related strains. The two extra introns in KL14-4A have been mapped between the Eco RI site (72.4 units) and Hind1 site (71 units), the region where the two strains differ by 3.4 kbp. Assuming that the general pattern of restriction sites is an index of overall sequence homology, the rest of the gene appears to have a similar organization in both strains. As shown in Fig. 9, the extra 3.4 kbp in KL14-4A can be accounted for by the two introns located between the Eco RI and Hind1 sites.

DISCUSSION
In the accompanying article (l), we described the properties of a set of p -clones with different genetic markers in the cytochrome b gene. The clones were characterized with respect to the physical properties of their mtDNAs and were used to construct a detailed restriction map of the cytochrome b region in S. cereuzsiue D273-10B. In the present study, the clones have been further exploited for the purpose of sequencing the cytochrome b gene.
The nucleotide sequence of yeast mtDNA has been obtained from 71.4 to 80.2 units. This span of the genome contains all the cytochrome b markers represented by our collection of mit-mutants. The sequence includes three coding regions that are proposed to be the exons of the yeast cytochrome b gene. The three exons generate a protein that is 50% homologous to bovine cytochrome b based on the sequence of the bovine gene.3 The fist exon codes for 252 to 253 residues from the NHZterminal end of the protein. The second exon consists of a very short reading frame that codes at most for 18 amino acids. The last 115 to 116 amino acids from the COOHterminal end are encoded by the third exon. The entire protein is 385 amino acids long and from its composition has a molecular weight of 44,000. This value is somewhat higher than previous estimates of 30,000 to 32,000 obtained by polyacrylamide gel electrophoresis (11,12). It has been pointed out, however, that cytochrome b behaves anomalously in sodium dodecyl sulfate gel electrophoretic systems and the value of 30,000 is probably a serious underestimation of the true size (26,27).
The cytochrome b gene occupies the wild type map from 71.5 to 76.1 units. In addition to the three exons, the gene has two intervening sequences that are 1414 and 733 nucleotides long, respectively. The first intron is especially interesting since it contains a reading frame that is continuous with the first exon and is potentially capable of coding for a second protein with 384 amino acids. Whether this sequence in fact corresponds to a separate gene is not known at present. There are several features of the sequence that tend to argue against a coding function. Most yeast mitochondrial genes sequenced to date contain 27 to 31% G+C (16-20). In contrast, the G+C content of the intron sequence is only 23%. The amino acid composition of the intron product is also unusual, being very rich in asparagine, lysine, tyrosine, and other amino acids whose codons do not have a G or C. Finally, the reading frame makes use of certain codons such as the CGN family for arginine and UGG for tryptophan that have not appeared in any of the identified yeast mitochondrial genes. Despite these reservations, we cannot exclude the possibility that the sequence might serve some real function, perhaps by coding for a protein necessary for processing of the cytochrome b messenger. This possibility is currently being investigated.
The sequence of the second intron has an A+T-rich character. This intron also has a short sequence with a high G+C content and several Hpa I1 and Hue 111 sites. Part of this ' Drs. F. Sanger and B. Barrell, private communication.

DNA Sequence and
Organization of the Cyt b Gene 9837 cluster is inversely homologous to a 14-nucleotide sequence further upstream in the intron. These inverted repeated sequences may be involved in some initial processing event leading to the removal of part of the intron sequence.
Although the yeast cytochrome b gene has been found to have a mosaic structure in all the yeast strains examined, the number of coding elements appears to vary depending on the strain. S. cereuisiae D273-10B has a relatively short mitochondrial genome, lacking a 3.4-kbp insert in the cytochrome b region. This extra 3.4-kbp sequence is present in most other yeast strains (8-10,25). Van Ommen et al. (25) have presented evidence that the 3.4-kbp insert occurs in a region of the cytochrome b gene where there are two long introns of approximately 2 and 1.4 kbp. The structure of the cytochrome b gene in D273-10B can be easily reconciled with the five-exon and three-intron structure proposed for other yeast strains if it is assumed that exon B1 is further fragmented into three smaller exons. In fact, the sizes of the fiist three cytochrome b exons in KL14-4A measured in RNA-DNA hybrids (25) agree reasonably well with the single b l exon of D273-10B reported here.
The definition at the DNA sequence level of the exon and intron regions makes it possible to study the location of mitochondrial mutations that abolish the synthesis of cytochrome b. The DNA sequences of the low complexity pclones have already provided some information about the location of 10 different cytochrome b mutations from our collection of mit-mutants. Based on the genotypes of the clones sequenced, all the cob2 mutations have been found to fall within the fist cytochrome b exon. The cobl mutants, however, represent a more heterogeneous group. While some cobl mutants appear to be in the last exon (b3), other mutants map either in the short middle exon or in the flanking introns. At least one mutation (MS-ZOO) falls completely inside of the intron sequence between b2 and b3. Such strains will be useful in future studies on the mechanism by which the exons are spliced in the messenger RNA.
Cytochrome b is an important electron transfer carrier of the mitochondrial respiratory chain. Spectral and kinetic evidence indicate the presence of two distinct b type cytochromes ( b k and bt) (28,29). The question of whether these are physically different proteins has never been resolved. Most of the genetic data, however, are consistent with the existence of a single cytochrome b gene. This has been c o n f i e d by the sequence of the gene reported here. Furthermore, the primary structure of the yeast cytochrome b now allows the chemical basis for the two spectral species to be tested experimentally.